Parallelise our CI/CD deployment of hubs

[sgibson91]

fixes #818

Summary

This PR implements a first pass of parallelised deployment of hubs in our CI/CD as discussed in #1131 There is still work to be done to make this match the vision in that issue, but I think this is a good start and we can iterate in future PRs.

What has changed

• I have renamed .github/actions/deploy to .github/actions/setup-deploy. It now contains all the steps required to be ready to deploy a hub to a cluster, many of which we had originally defined .github/workflows/deploy-hubs.yaml, but no longer performs the deploy itself.
• Parallelised jobs have been added to .github/workflows/deploy-hubs.yaml.
  • The generate-jobs job generates all the matrix jobs required as defined by the deployer logic implemented in (Attempt 2) Add logic to the deployer to decide which clusters/hubs require a helm upgrade #1112. In PRs, this job provides a rich output table of jobs.
  • The upgrade-support-and-staging job upgrades the support chart and/or the staging hub on each cluster that require it in parallel. This job requires generate-jobs to complete before running.
  • The upgrade-prod-hubs job then upgrades every production hub that requires it in parallel. This job requires both generate-jobs and upgrade-support-and-staging to complete before running.

Failing test

The message appearing in the failed test is being tracked here: jitterbit/get-changed-files#42

But I actually think this will resolve itself once this PR is merged. I opened a test PR to my test repo via a fork and all worked fine: sgibson91/helm-upgrade-decision-logic#6

[sgibson91]

I don't think the prettier pre-commit failure is related to what I've changed here, but it's really hard to tell from the logs. There's no spaces between the filenames and so it's just one giant line - very difficult to parse 😕

[GeorgianaElena]

This work that has been done around parallelizing hub deployment is just awesome 🚀 ❤️
Thanks @sgibson91!
I left a few questions, but I believe we should merge this and see how it goes. However, I believe that others on the @2i2c-org/tech-team should should also have a look before since it's a pretty significant change.

[consideRatio]

Nice work @sgibson91 this LGTM!!!

I've added a few comments focused on documentation that I think are quite important for maintenance sustainability but not technically critical - from the technical standpoint is approved as it is!

[sgibson91]

Note: we may need to address #1171 first

[sgibson91]

@consideRatio I propose we start our co-working tomorrow by hitting merge on this, fix any bugs that happen to arise, and then decide where to go with the rest of our time from there

[consideRatio]

Wieee! London collaboration merge!